OpenAI ships ChatGPT Images 2.5 with API Flare and Sunburst

OpenAI is shipping ChatGPT Images 2.5. On 8 September 2026 the lab said the ChatGPT image model is rolling out to all ChatGPT, ChatGPT Work, and Codex users on desktop, mobile, and web. The API split is GPT-Image-2.5 Flare (fast default) and GPT-Image-2.5 Sunburst (higher precision, longer generation).

From the 8 September 2026 @OpenAI post: "ChatGPT Images 2.5 is rolling out today to all ChatGPT, ChatGPT Work, and Codex users across desktop, mobile, and web."

The OpenAI product post calls Images 2.5 its new state-of-the-art image model and reports generation latency down by up to 50% compared with Images 2.0. It also claims more natural lighting and richer textures, better subject preservation from reference photos, and more reliable multi-turn edits.

ChatGPT product additions named in that post include Sketch (type "@Sketch" or open chatgpt.com/sketch), templates such as Poster and Merch, comments placed on images, and optional prompt sharing when you share an image.

On the API, GPT-Image-2.5 Flare (`gpt-image-2.5-flare`, snapshot `gpt-image-2.5-flare-2026-09-08`) is the default. The product post says higher-quality images than GPT-Image-2 at 50% lower latency. GPT-Image-2.5 Sunburst (`gpt-image-2.5-sunburst`, snapshot `gpt-image-2.5-sunburst-2026-09-08`) is the slower precision option. Both model cards list quality settings including xhigh and max, and token rates of $5 per million text input tokens, $8 per million image input tokens, and $30 per million image output tokens. See API pricing.

@OpenAIDevs on X said: "Flare delivers higher-quality images at 50% lower latency than GPT-Image-2. Sunburst adds even more precision for detailed editing. More natural lighting, richer textures, and better handling of complex layouts and transparent backgrounds."

OpenAI points to a ChatGPT Images 2.5 system card and says it still uses C2PA metadata and invisible watermarking.

Key facts: - 8 September 2026: ChatGPT, ChatGPT Work, and Codex rollout (OpenAI). - API models: Flare (fast default) and Sunburst (precision, longer generation). - Latency "up to 50%" vs Images 2.0, and "50% lower" vs GPT-Image-2, is OpenAI-reported. - Sketch, templates, on-image comments, and prompt sharing are ChatGPT product features. - Safety writeup: system card; C2PA plus watermark (vendor).

How to read: the hard facts are the named surfaces, model ids, and the Flare/Sunburst split. Quality, latency, and "state-of-the-art" lines stay OpenAI claims. Partner blurbs in the product post (Adobe, Runway, Manus, Higgsfield) are vendor quotes.

Independence: Frontier Flux is independent of OpenAI. The X post is the alert; the product post, API cards, and system card are the evidence.